Quick Start

Start here

1. Get an API key.
2. Copy the request below.
3. Confirm you got a markdown response back.

If you reach step 3, you are unblocked for almost every other page in this docs set.

Authentication

Create an account at fastcrw.com/register, then send the key in the Authorization header:

Authorization: Bearer YOUR_API_KEY

First request

curl -X POST https://fastcrw.com/api/v1/scrape \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://example.com",
    "formats": ["markdown"]
  }'

First response

{
  "success": true,
  "data": {
    "markdown": "# Example Domain\n\nThis domain is for use in illustrative examples...",
    "metadata": {
      "title": "Example Domain",
      "sourceURL": "https://example.com",
      "statusCode": 200,
      "elapsedMs": 32
    }
  }
}

That is the default CRW shape most users should start with: one URL, one markdown output, no extra knobs.

The same request in code

:::tabs ::tab{title="cURL"}

curl -X POST https://fastcrw.com/api/v1/scrape \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url":"https://example.com","formats":["markdown"]}'

::tab{title="Python"}

import requests

resp = requests.post(
    "https://fastcrw.com/api/v1/scrape",
    headers={
        "Authorization": "Bearer YOUR_API_KEY",
        "Content-Type": "application/json",
    },
    json={
        "url": "https://example.com",
        "formats": ["markdown"],
    },
)

print(resp.json()["data"]["markdown"])

::tab{title="Node.js"}

const resp = await fetch("https://fastcrw.com/api/v1/scrape", {
  method: "POST",
  headers: {
    "Authorization": "Bearer YOUR_API_KEY",
    "Content-Type": "application/json"
  },
  body: JSON.stringify({
    url: "https://example.com",
    formats: ["markdown"]
  })
});

const body = await resp.json();
console.log(body.data.markdown);

:::

Pick the next page

:::cards ::card{icon="code" title="Need a known URL?" href="#scraping" description="Stay with scrape and add formats, selectors, JS rendering, or extraction."} ::card{icon="search" title="Need unknown URLs?" href="#search" description="Use search first, then scrape only the results you care about."} ::card{icon="map" title="Need discovery on one site?" href="#map" description="Use map when you need reachability before you recurse."} ::card{icon="globe" title="Need multiple pages?" href="#crawling" description="Use crawl for bounded recursion after you validate the target section."} :::

MCP path

If your real goal is to give an AI tool live web access, go straight to MCP:

claude mcp add crw -- npx crw-mcp

For Codex, Cursor, Windsurf, and others, continue in MCP Server.

Self-host path

If you want a local or private deployment instead of the hosted API:

docker run -p 3000:3000 ghcr.io/us/crw

Then call the local API:

curl -X POST http://localhost:3000/v1/scrape \
  -H "Content-Type: application/json" \
  -d '{"url":"https://example.com","formats":["markdown"]}'

Use Self-Hosting for the full deployment path and Installation for package-level install options.

Common mistakes

• Using too many options in the first request. Start with formats: ["markdown"].
• Turning on JS rendering before checking whether plain HTTP already works.
• Jumping into crawl before validating the target with scrape.
• Treating search as self-hosted. In these docs, search is the hosted/cloud path unless noted otherwise.

What to read next

• Scrape for the canonical single-page flow
• Search for discovery-first workflows
• Authentication for key handling
• Self-Hosting if you want to move off the hosted path